Engineering Structures 94 (2015) 54–69

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Structural performance of confined masonry walls retrofitted using

ferrocement and GFRP under in-plane cyclic loading
Mosaad El-Diasity a,⇑, Hussein Okail b, Osama Kamal a, Mohamed Said a
a
Structural Engineering Department, Shoubra Faculty of Engineering, Benha University, Egypt
b
Structural Engineering Department, Ain Shams University, Egypt

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents the results of in-plane cyclic loading tests conducted on confined masonry walls ret-
Received 2 May 2014 rofitted using low-cost ferrocement and GFRP systems. Ten wall assemblies with a 0.80-scale were built,
Revised 13 March 2015 consisting of a clay masonry panel, two confining columns and a tie beam. The assemblies were tested
Accepted 14 March 2015
under a combination of a vertical load and lateral reversed cyclic loading with a displacement controlled
loading protocol up to failure. Wall panels had various configurations, namely, solid walls, perforated
walls with window and door openings. Two composite materials (ferrocement and GFRP) and three retro-
Keywords:
fitting configurations (diagonal ‘‘X’’, corner, and full coverage) were investigated. Key experimental
Confined masonry
Seismic behavior
results showed that the proposed upgrading techniques improved the lateral resistance of the confined
Cyclic loading walls by a factor ranging from 25% to 32%with a significant increase in the ductility and energy absorp-
Shear failure tion of the panel ranging from 33% to 85%; however, the improvement in lateral drifts was less significant.
Ferrocement Regarding the upgrading configurations, the diagonal ‘‘X’’ and full coverage can help prevent diagonal
Expanded mesh shear failure especially in tie columns and convert the failure mode to a panel-rocking mode.
Glass woven fibers Additionally, in all retrofitting cases, collapse was significantly delayed by maintaining the wall integrity
under large lateral deformations. A good agreement was found by comparing deformed shapes, crack pat-
terns and capacity curves of finite element models included in this study.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Background information
Confined masonry construction has emerged as a building tech-
nology that offers an alternative to both unreinforced masonry and
infilled reinforced concrete (RC) frame construction. The system
consists of masonry panel confined with horizontal and vertical
RC elements with light reinforcement ratio. The confined masonry
is considerably different from infilled RC frame with respect to its
construction methodology, as the masonry wall is laid before the
tie columns (Brzev [1]). The system also differs by the load transfer
mechanism under gravity and lateral load, the masonry wall trans-
mit the gravity load from the slab above down to the foundation.
The walls act as bracing panels; confining elements (tie column
and tie beam) provide restraint to masonry panel and protect it
from complete disintegration especially under large lateral defor-
mations (Yoshimura et al. [2]).
The in-plane response of a confined masonry wall is distinctly
different from that of reinforced concrete infilled frames.
⇑ Corresponding author.
E-mail address: ec_cai_mosaad.hassan@yahoo.com (M. El-Diasity).
Although the confined masonry wall experiences both flexural
and shearing deformations, the masonry infill deforms in a shear
mode within a frame that attempts to deform in flexure, resulting
in separation of the frame and infill wall along the tension diagonal
(Riahi et al. [3]).
As a result, improvement in deformation and energy dissipation
characteristics are also achieved if the system is properly con-
structed (Tomazevic and Klemenc [4]).
A confined masonry wall subjected to in-plane lateral earth-
quake loading develops either a shear or flexural failure mecha-
nisms (Tomazevic and Klemenc [4]; Yoshimura et al. [2]). Shear
failure mechanism is characterized by distributed diagonal crack-
ing in the wall. The damage is caused either by the bond destruc-
tion at the mortar–brick interface (shear-friction mechanism), or
tensile cracking in the masonry units (Tomazevic [5]). Flexural fail-
ure mechanism due to in-plane lateral loads is characterized by
horizontal cracking of the bed-joints located on the tension side
of the wall (Yoshimura et al. [2]). Separation of the tie-columns
from the wall was observed in some cases when a toothed wall-
to-column connection was absent, and there were no connecting
ties between the tie-column and the wall. Flexural mechanism is
not as critical as shear mechanism since it does not lead to brittle

http://dx.doi.org/10.1016/j.engstruct.2015.03.035
0141-0296/Ó 2015 Elsevier Ltd. All rights reserved.
 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69 55

Table 1
Design summary of tested walls.

Wall ID Wall state Masonry panel Retrofitting configuration Long RFT Trans RFT Vertical stress (MPa)
Un-retrofitted
CLY-S-CTRL Unretrofitted Solid – 4T10 T6@20 cm 0.5
CLY-P-W Unretrofitted Perforated window – 4T10 T6@20 cm 0.5
CLT-P-D Unretrofitted Perforated door – 4T10 T6@20 cm 0.5
Retrofitted
CLY-D-FERRO Retrofitted Door perforated Ferrocement full coverage 4T10 T6@20 cm 0.5
CLY-W-FERRO Retrofitted Perforated window Ferrocement full coverage 4T10 T6@20 cm 0.5
CLY-S-FERRO-C Retrofitted Solid Ferrocement connections 4T10 T6@20 cm 0.5
CLY-S-FERRO-X Retrofitted Solid Ferrocement X-diagonal 4T10 T6@20 cm 0.5
CLY-S-FERRO-F Retrofitted Solid Ferrocement full coverage 4T10 T6@20 cm 0.5
CLY-S-GFRP-X Retrofitted Solid GFRP X-diagonal 4T10 T6@20 cm 0.5
CLY-S-GFRP-F Retrofitted Solid GFRP full coverage 4T10 T6@20 cm 1.0

Tested Walls

Fig. 1. Typical details of the tested walls assemblies.

(a) Footing with column reinforcement (b) Masonry wall construction

(c) Concrete casting of tie column (d) Concrete casting of tie Beam
Fig. 2. Construction sequence of a typical wall assembly.
56 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69

(a) Spatter-dashing of wall (b) Fixing the expanded Mesh (c) Laying of mortar layer
Fig. 3. Procedures of retrofitting walls by ferrocement.

Table 2
Properties of the GFRP material.
(a)
Commercial name Weight Breaking Strength Fiber diameter (lm) (b)
EWR-600 600 g/m2 3850 N/50 mm 13
(c)

failure, although crushing and disintegration of masonry in the

compression toe area of the wall may take place (Tomazevic [5]).
Moreover, during an earthquake, the masonry panels are sub-
jected to in-plane and out-of-plane loads simultaneously. The
out-of-plane load-carrying capacity of these masonry panels may (d)
be substantially weakened after being damaged, endangering their
overall safety and stability. The extent of damage and likelihood of
wall collapse in the out-of-plane direction also depends on the
type of floor diaphragm (rigid or flexible), the spacing between
confining elements, and the connection of masonry panel with
(c)
adjacent confining elements. Good bonding between the masonry
wall and adjacent RC tie-columns is essential for satisfactory earth-
quake performance, and for delaying undesirable cracking and sep-
aration of wall with confining elements. The shake table test on Fig. 5. Test set up of diagonal compression test.
confined masonry walls conducted by Tu et al. [6] concluded that
the strong boundary connection prevent masonry panel from fall-
A prediction method on structural performance for confined
ing out of the frame and thus can sustain considerable out-of-plane
masonry walls and for retrofitting of confined masonry walls were
seismic loads.
presented by Lourdes Ana et al. [10] with comparison of results of
Opening size and the degree of coupling affect both initial stiff-
structural test performed at structural laboratory in CISMID UNI-
ness and cracking pattern. The rate, at which stiffness degrades,
Lima-Peru. The cyclic loading tests on walls were carried out two
however, is almost independent of these factors (Ishibashi et al.
different kinds of the retrofitted schemes of confined masonry
[7]). While excessively large openings could reduce shear capacity
walls: the one was a wire mesh wrapping alternative and the other
of confined masonry walls by almost 50% (Gostic and Zarnic [8]),
was the combined materials such as adhesive and disposal mats
their effect on seismic performance is almost negligible when size
with different arrangements. The methodology to evaluate struc-
is restrained to approximately 10% of the wall gross area (Yanez
tural performance for retrofitted confined masonry wall was
et al. [9]).

(a) Laying of thin layer of mortar (b) Painting polyester risen (c) Laying of GFRP sheets
Fig. 4. Procedures of retrofitting walls by GFRP.
 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69 57

a – Hydraulic jack; b – Load cell; c – Loading shoes; d – Masonry specimen

(a) Wallet W1 (b) Wallet W2

Fig. 6. Failure patterns of the tested wallets.

Vertical Hydraulic Jack

Horizontal Hydraulic Jack

Footings fixation system

(Bolts and Footing fixation system

horizontal jack) Bolts and thickned steel plats

Fig. 7. Test setup, boundary conditions and loading mechanism.

Fig. 8. Cyclic displacement protocol.

suggested. And the accuracy of the prediction method could be
verified, and the applicability of its model in analysis was con-
firmed using pushover analysis.
A mathematical model was developed by Flores and Alcocer
[11] to represent the non-linear behavior of confined masonry
structures built with hand-made clay bricks. The suggested
Fig. 9. Instrumentation scheme.
envelope curve of the model is a tri-linear force–deformation
curve, which is calculated from material properties and wall
geometry. The parameters for envelope curve are detailed accord-
ing to masonry walls with and without horizontal reinforcement at
58 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69
Fig. 10. Crack pattern for CLY-S-CTRL.
Fig. 11. Crack pattern for CLY-S-FERRO-C.
Fig. 12. Crack pattern for CLY-S-FERRO-X with rocking at connection.
Fig. 13. Crack pattern for CLY-S-GFRP-X with rocking at connection.
bed joints. Also an experimental study on effects of height of lateral
forces, column reinforcement and wall reinforcements on seismic
behavior of confined masonry was developed by Yoshimura et al.
[12] to study the ultimate shear strengths of the confined concrete
hollow block masonry specimens with developed equation.
ElGawady et al. [13] conducted dynamic in-plane tests on six
slender un-reinforced masonry walls retrofitted using different
types of FRP and were subjected to a series of simulated earth-
quake motions on an earthquake simulator. It was observed that
FRP upgrading improve the wall lateral resistance by a factor rang-
ing from 1.3 to 2.9 and the X shape upgrading configuration had
the maximum drift of all specimens.
Grids of GFRP were used by Yu et al. [14] to reinforce a fast set-
ting polyurea spray to strengthen eleven URM walls subjected to
in-plane diagonal compressive load and it was concluded that
the failure mode was directly affected by strengthening schemes
(orientations) and the lateral load capacity increased by a factor
of 1.1–1.43 for brick walls and 1.1–1.6 for concrete block walls.
As a part of a research program aiming at developing struc-
turally and economically efficient hybrid building system for
developing countries in general and for Egypt in particular. This
paper investigates the lateral load behavior of retrofitted confined
masonry walls using low cost ferrocement and GFRP systems
designed and built using locally available materials and with com-
mon workmanship and construction practice.
2. Experimental program
The experimental program investigates the effectiveness of
composite materials; namely ferrocement using expanded mesh
and sheets of glass woven textile fibers (GFRP), as externally
bonded upgrading materials for the in-plane retrofitting of CM
walls. The two alternatives represent low-cost retrofitting
methods. The experimental program includes testing both
 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69 59

Fig. 14. Crack pattern for CLY-S-FERRO-F with enlarged rocking at connection.

Fig. 15. Crack pattern for CLY-S-GFRP-F with enlarged rocking at connection.

Fig. 16. Crack pattern for CLY-P-W. Fig. 17. Crack pattern for CLY-W-FERRO.

un-retrofitted and retrofitted CM wall assemblies up to failure
under reversed cyclic lateral loads.
2.1. Description of the tested walls
Ten wall assemblies with a 0.80-scale were built, using 0.80-
scale clay brick units. The assemblies consisted of a clay masonry
panel, two confining columns and a tie beam. The assemblies were
tested under a combination of a constant vertical load and lateral
cyclic loading with displacement controlled loading protocol up
to failure. Table 1 summarizes the tested walls. The walls were
selected such that it covers the range of different perforations,
namely solid walls or walls with window or door openings. The
test matrix investigates the use of ferrocement or GFRP in
retrofitting these alternatives using multiple arrangements.
Coverage of the walls was done either by fully covering the entire
wall with the confining elements, covering the corners of the panel
with an overlap with the confining elements or by forming a diago-
nal X-brace in the masonry panel. Typical details of tested walls are
shown in Fig. 1. The construction of wall assemblies is shown in
Fig. 2a–d, the thickness of wall for all assemblies was 200 mm.
Five walls were retrofitted using one ferrocement layer consist-
ing of expanded wire mesh with a grid size of 15  35 mm and
thickness of 1.6 mm, the mesh was from mild steel (Grade 240/
350) with yield stress (fy) of 240 MPa, and ultimate tensile strength
(fu) of 350 MPa and modulus of elasticity (Es) of 200 GPa. The
expanded mesh was fixed to the masonry wall after spatter-
dashing it by nails every 100 mm in both directions, then covering
the expanded mesh with a mortar of 20 mm thickness as shown in
60 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69
Fig. 18. Crack pattern for CLY-P-D.
Fig. 19. Crack pattern for CLY-D-FERRO.
Fig. 3. The mortar characteristics (3 sand: 1 cement & w/c ratio of
0.5). The 28-days characteristic compressive strength of mortar,
(fcu), was on average 21.8 MPa.
The assemblies CLY-S-FERRO-C and CLY-S-FERRO-X were cov-
ered by two perpendicular expanded meshes for each face with
mortar thickness of 30 mm.
Two walls were retrofitted using GFRP sheets. Firstly a thin
layer of mortar was applied on the masonry panel to make the sur-
face of as plane as possible, then the surface was painted by a layer
of polyester resin and the GFRP sheets with properties as men-
tioned in Table 2. Fig. 4 shows the retrofitting procedures.
For control purposes, standard concrete cubes were cast along-
side the walls and were tested at the same day as the walls, the 28-
days concrete characteristic compressive strength, (fcu), was on
average 25.4 MPa. Standard five brick masonry prisms were built
next to the walls and tested at the same day of the wall testing.
0
The mean compressive strength of the clay masonry prisms (f m )
was 5.7 MPa. The longitudinal reinforcement of beams and col-
umns was made of deformed steel bars (Grade 360/520) of yield
stress (fy) of 360 MPa, and ultimate tensile strength (fu) of
520 MPa. The transverse reinforcement was made of mild steel
smooth bars (Grade 240/350) of yield stress (fy) of 240 MPa, and
ultimate tensile strength (fu) of 350 MPa. All the previous
reinforcement had a modulus of elasticity (Es) of 200 GPa. The
walls were left to cure for 28 days before testing and were white
washed with non-latex paint to ease the visualization of the devel-
oped cracks during tests.
Two masonry wallets with size (850  850  95 mm) were built
next to the walls. The wallets tested by diagonal compression test
as per ASTM E519-02 [15] to determine the diagonal tensile
strength; a schematic test setup of diagonal compression test is
shown in Fig. 5.
The test setup is composed of two steel loading shoes, which
were fixed on the two opposite corners on the wallet. The load is
applied to the wallet by a hydraulic jack on the steel loading shoes
positioned on top of the specimen and transmitted to the other
shoe which was placed at the bottom of the specimen.
The test was introduced to simulate a pure shear stress state,
the ASTM [15] assumes that the diagonal compression test pro-
duces a uniform shear stress and a Mohr’s circle centered in the
origin of the Cartesian system of axis. In that case the value of
the shear stress s is equal to the principal tensile stress ft. The shear
stress s is obtained by Eq. (1):
0:707 P
s¼ ð1Þ
An
where P is the load applied by the jack and An is the net area of the
specimen, calculated as follows:
 
wþh
An ¼ tn ð2Þ
2
where w is the specimen width, h is the specimen height, t is the
thickness of the specimen and n is the percentage of the unit’s gross
area that is solid, expressed as a decimal. In the present work the
value n = 0.84 was adopted. The same expressions are also pre-
sented in Eurocode 6 [16].
The RILEM committee [17] considers that the stress field is not
uniform and proposes the following expression to evaluate the ten-
sile strength of masonry:
0:5 P
ft ¼ ð3Þ
An
Fig. 6a and b shows the failure and crack patterns for both wal-
lets W1 and W2, both of them shared the same mode of failure
characterized by direct diagonal tension crack under failure loads
of 105 and 111 kN respectively.
According to Eq. (3), the tensile stress ft for specimens equals
0.77 and 0.81 MPa respectively.
2.2. Test setup, boundary conditions and loading scheme
The walls were tested up to failure under a combined constant
vertical load and in-plane cyclic lateral load, Fig. 7 shows the test
setup of the walls. In this respect, a single concentrated load of
250 kN (for all walls except 500 kN for specimen CLY-S-GFRP-F)
was firstly distributed by a stiff steel distributor I-beam laid on
top of secondary steel beam and separated by four rolling steel
cylinders as shown in Fig. 7. The secondary steel beam was laid
on top of the concrete tie beam of the assembly using gypsum
bedding to avoid stress concentration. The purpose of the rolling
cylinders is to allow the wall to displace laterally while maintain-
ing the distributed vertical load. The load was chosen to simulate
that of a typical module in a five-story residential building with
commonly used module dimensions. The lateral cyclic load was
applied using a 900 kN hydraulic actuator. The horizontal action
is applied to the wall via control displacement at a rate of
60 lm/s, full displacement protocol was programmed for each
amplitude increment aiming at strength and degradation
assessment as shown in Fig. 8. The footing was fixed to the reaction
floor by two strong pre-tensile bolts (Ø50 mm) spaced by 3.0 m to
prevent overturning and sliding of the footing during test, in addi-
tion, a horizontal hydraulic jack was placed horizontally at footing
level and reacting against a steel reaction column, in the opposite
 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69
(a) CLY-S-CTRL & CLY-S-FERRO-C
(c) CLY-S-CTRL & CLY-S-FERRO-F
Fig. 20. Hysteretic curves for retrofitted and un-retrofitted solid walls.
side a thick steel plates were placed to restrain the sliding of
footing during the test as shown in Fig. 7. At the end of each cycle,
the displacement was held constant for a period of 2 min, during
which measurements, observations as well as marking of the visi-
ble cracks too place.
2.3. Instrumentation
Measurements were made for displacements, steel, and con-
crete strains at key locations of the tested walls, displacement were
measured using six 0.01 mm accuracy electrical linear variable dis-
tance transducers (LVDTs), coded (D), positioned as shown on
Fig. 9. The steel strain in the longitudinal and transverse reinforce-
ment was monitored using five electrical strain gauges of 10 mm
gauge length and 120 Ohm resistance, coded (S), as shown in
Fig. 9. All LDVTs and strain gauges were connected to a computer
controlled data acquisition system. The crack patterns was con-
tinuously monitored and printed on the walls with the associated
displacement level printed next to it.
61
(b) CLY-S-CTRL & CLY-S-FERRO-X
(d) CLY-S-CTRL & CLY-S-GFRP-X
3. Experimental observations and discussion
3.1. Failure pattern
Fig. 10 shows the failure and crack pattern for the un-retrofitted
wall (CLY-S-CTRL) the mode of failure may be characterized by
shear failure of the confining columns and diagonal bed-joint
cracking of the masonry panel. It is worth nothing that no sep-
aration was observed at the toothed interface between the confin-
ing columns and the masonry panel, clearly confirming a
significant difference between infilled frames and confined
masonry panel.
Fig. 11 shows the failure and crack pattern for the retrofitted
solid wall (CLY-S-FERRO-C), it can be clearly seen that the presence
of the ferrocement layers at corners delay the diagonal shear cracks
in masonry panel to propagate in confining column which occurred
in the un-retrofitted wall, later the failure was due to diagonal
shear cracks developing around the retrofitted corners. The failure
occurred at increased capacity and ductility.
62 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69

(a) CLY-P-W& CLY-W-FERRO (b) CLY-P-D& CLY-D-FERRO

Fig. 21. Hysteretic curves for retrofitted and un-retrofitted perforated walls.

Fig. 22. Hysteretic curve for wall CLY-S-GFRP-F. Fig. 23. Envelope curves for solid walls.

Fig. 12 shows the failure and crack pattern for the retrofitted
solid wall (CLY-S-FERRO-X), it can be clearly seen that the presence
of the ferrocement layers on the path of the principle tension fields
prevented the diagonal shear cracks in masonry panel, later some
minor flexural cracks were seen in the confining columns, the fail-
ure was due to rocking for the undamaged panel after the slightly
reinforced confining column reached its ultimate tension capacity
at base. The failure mode of the retrofitted wall (CLY-S-GFRP-X)
was similar to specimen (CLY-S-FERRO-X) as shown in Fig. 13.
For the retrofitted solid wall (CLY-S-FERRO-F), the presence of a
full coverage with ferrocement layers prevented any cracks in
masonry panel and the confining elements. The failure was due
to rocking for the un-cracked panel as shown in Fig. 14. The failure
mode of the retrofitted wall (CLY-S-GFRP-F) was similar to speci-
men (CLY-S-FERRO-F) as shown in Fig. 15.
Figs. 16 and 17 clarify the difference of failure mode for both
un-retrofitted and retrofitted walls with window opening. The
presences of ferrocement layer prevent shear failure of confining Fig. 24. Envelope curves for walls with window.
 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69 63

Fig. 26. Calculation of energy dissipation.

Fig. 25. Envelope curves for walls with door.
load was 383 kN as opposed to 304 kN for the reference un-retro-
columns and minimized the cracks in the masonry panel. The fail-
fitted wall (CLY-S-CTRL). This corresponding to about 26% increase
ure occurred at increased capacity and ductility as will be
in the lateral load resistance of the wall. Fig. 20a shows the hys-
described in details in a later section. It can be clearly noticed that
teretic curves for both assemblies.
the shear failure in the confining elements was shifted to diagonal
Wall assembly CLY-S-FERRO-X was retrofitted using two layers
cracks in the masonry panel around the window opening.
of ferrocement at diagonals on both sides. The wall’s ultimate load
The failure pattern of the un-retrofitted wall with door opening
was 387 kN as opposed to 304 kN for the reference un-retrofitted
may be characterized by shear failure with diagonal struts forming
wall (CLY-S-CTRL). This corresponding to about 27.5% increase in
in the two piers, as shown in Fig. 18, the presence of ferrocement
the lateral load resistance of the wall. Fig. 20b shows the hysteretic
layer prevented these diagonal shear cracks in piers, later at higher
curves for both assemblies.
levels of lateral loads the failure was due to the rocking of the piers
Wall assembly CLY-S-FERRO-F was retrofitted using one layer of
and the door lintel due to the formation of plastic hinges at their
ferrocement covering full masonry panel on both sides. The wall’s
ends, as shown in Fig. 19.
ultimate load was 377 kN as opposed to 304 kN for the reference
un-retrofitted wall (CLY-S-CTRL). This corresponding to about
3.2. Behavior of retrofitted walls 24% increase in the lateral load resistance of the wall. Fig. 20c
shows the hysteretic curves for both assemblies.
Wall assembly CLY-S-FERRO-C was retrofitted using two layers Wall assembly CLY-S-GFRP-X was retrofitted two layers of GFRP
of ferrocement at connections on both sides. The wall’s ultimate sheets at diagonals on both sides. The wall’s ultimate load was

Table 3
Summary of test results.

Wall ID Direction Maximum load Displacement relative to Cracking load Maximum displacement
maximum load
Load (kN) Variation% Drift (mm) Variation% Load (kN) Variation% Drift (mm) Variation%
CLY-S-CTRL Push 300 – 7 – 250 – 16 –
Pull 280 – 10 – 170 – 14.8 –
CLY-S-FERRO-C Push 380 27a 8 14a 340 36a 16.2 1a
Pull 350 25a 7 30a 350 106a 14 5a
CLY-S-FERRO-X Push 387 29a 10.5 50a 330 32a 14.1 11a
Pull 370 32a 12 20a 340 100a 14 5a
CLY-S-GFRP-X Push 386 29a 11 57a 330 32a 13.8 14a
Pull 360 29a 10.5 5a 300 76a 12.2 17a
CLY-S-FERRO-F Push 377 26a 7 0.5a 377 51a 11.8 26a
Pull 375 34a 10 0.4a 375 121a 12.1 18a
CLY-S-GFRP-F Push 540 n/a 14 n/a 540 n/a 14 n/a
Pull 510 n/a 13 n/a 510 n/a 13 n/a
CLY-P-W Push 250 – 12 – 155 – 16.2 –
Pull 258 – 10 – 150 – 15 –
CLY-W-FERRO Push 335 34b 7.3 39b 225 45b 19.2 19b
Pull 315 22b 7 30b 230 53b 15 0.3b
CLT-P-D Push 205 – 15 – 140 – 27 –
Pull 209 – 20 – 120 – 22.5 –
c
CLY-D-FERRO Push 270 32 16 7c 210 50c 34 26 c

c
Pull 275 32 20 0.4c 145 21c 27.5 22 c

a
Percentage referenced to the specimen (CLY-S-CTRL).
b
Percentage referenced to the specimen (CLY-P-W).
c
Percentage referenced to the specimen (CLY-S-CTRL).
64 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69
(a) Solid walls
(c) Walls with door
Fig. 27. Cumulative energy dissipation for wall assemblies.
386 kN as opposed to 304 kN for the reference un-retrofitted wall
(CLY-S-CTRL). This corresponding to about 27% increase in the lat-
eral load resistance of the wall. Fig. 20d shows the hysteretic
curves for both assemblies.
Wall assembly CLY-W-FERRO was retrofitted using one layer of
ferrocement covering full masonry panel on both sides. The wall’s
ultimate load was 335 kN as opposed to 258 kN for the reference
un-retrofitted wall (CLY-P-W). This corresponding to about 30%
increase in the lateral load resistance of the wall. Fig. 21a shows
the hysteretic curves for both assemblies.
Wall assembly CLY-D-FERRO was retrofitted using one layer of
ferrocement covering full masonry panel on both sides. The wall’s
ultimate load was 275 kN as opposed to 209 kN for the reference
un-retrofitted wall (CLY-P-D). This corresponding to about 32%
increase in the lateral load resistance of the wall. Fig. 21b shows
the hysteretic curves for both assemblies. Envelope curves for solid
and perforated walls are shown in Figs. 23–25.
Wall assembly CLY-S-GFRP-F was retrofitted using one layer of
GFRP sheets covering full masonry panel on both sides and tested
(b) Walls with window
under constant vertical stress of 1.0 MPa. The wall’s ultimate load
was 540 kN. Fig. 22 shows the Hysteretic curve for the wall assem-
bly. Table 3 summarizes all test results for both un-retrofitted and
retrofitted walls.
3.3. Energy dissipation
Energy dissipation, Ed, through hysteresis damping is an impor-
tant aspect in seismic design response, Ed, has been represented, as
suggested by Hose and Seible [18], by area enclosed within the
force–displacement curve at each displacement level. This is the
horizontally-hatched area shown in Fig. 26. The vertically-hatched
region in the same figure represents the elastic strain energy, Es,
stored in an equivalent linear elastic system.
The cumulative energy dissipation at different displacement
levels of solid wall assemblies were presented in Fig. 27a. The fig-
ure showed that, an improvement in total energy dissipation of
about 47% has been achieved for the retrofitted assembly CLY-S-
FERRO-X corresponding to the control solid assembly CLY-S-CTRL.
 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69
(a) Solid walls
(c) Walls with door
Fig. 28. Hysteresis damping percentages for wall assemblies.
Fig. 27b shows the cumulative energy dissipation for assemblies
CLY-P-W AND CLAY-W-FERRO, an improvement in total energy
dissipation of about 85% was achieved.
An improvement of total energy dissipation of about 33% was
achieved for wall assembly CLY-P-D as shown in Fig. 27c.
3.4. Hysteretic damping
Hysteretic damping, nhyst, can be quantified based on the equal
area approach (Hose and seible [18]) that represents the same
amount of energy loss per loading cycle. The relationship between
the dissipated energy, Ed, the stored strain energy, Es, and the hys-
teretic damping is given by equation shown in Fig. 26.
The hysteretic damping was plotted against lateral top displace-
ment for solid and perforated walls are shown in Fig. 28a–c.
For the solid walls, the hysteretic damping ranges from 15% to
30% and the wall assembly CLY-S-FERRO-X was the largest per-
centage ranging from 22.5% to 32%.
An improvement in hysteretic damping was achieved for the
retrofitted perforated walls with windows and doors as shown in
Fig. 28b and c.
65
(b) Walls with window
3.5. Stiffness degradation
To assess the variation in wall stiffness with increased loading
and top displacement, the secant stiffness, defined as the ratio
between the lateral resistance and the corresponding top lateral
wall displacement, was used. The cycle stiffness of the specimen
at a certain displacement level was considered as the average of
stiffness in the positive and negative loading directions, as illus-
trated in Fig. 29a–c.
The trends of stiffness degradation for all walls were approxi-
mately similar and showed significant decreases with increased
top displacement.
4. Numerical analysis
The aim of this section is to establish a simple three-dimen-
sional nonlinear model for the tested wall assemblies that are cap-
able of capturing the key response features of the failure mode
shapes and crack patterns for each assembly and comparing it with
experimental results and with respect to the past references.
66 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69
(a) Solid walls
(c) Walls with door
Fig. 29. Stiffness degradation for wall assemblies.
In the past, numerous studies were conducted on the finite ele-
ment models of masonry walls as well as RC infilled frames.
Kaushik et al. [19] studied the uniaxial monotonic compressive
stress–strain behavior and estimated the modulus of elasticity of
bricks, mortar, and masonry as 300, 200, and 550 times their com-
pressive strengths, respectively.
There are two methods of finite element modeling for masonry
structures which consist of bricks and mortar: separated modeling
and integrated modeling. The former simulates brick and mortar
separately while the later simulates them integrity. For the sepa-
rated modeling approach, there are two categories: the first
assumes that brick and mortar are well integrated, and the element
nodes on the contact surface satisfy continuous displacement con-
dition. Hence, the degrees of freedom on the corresponding nodes
on the contact surface are coupled together. The other one consid-
ers bond-slip between brick and mortar, which requires the intro-
duction of interface elements (Huang et al. [20]).
(b) Walls with window
Analysis on the characteristics and features of masonry using
Solid65 elements in ‘‘ANSYSÒ [21]’’ was studied by Huang et al.
[20] with the shear property of joints in masonry structures under
different vertical load (r/fm) was numerically simulated.
Comparing the experimental results with the numerical ones, the
proposed values for the shear transfer coefficients for open and
closed crack of Solid65 elements for simulating masonry structures
were suggested.
4.1. Finite element model
The non-linear finite elements analysis was carried out using a
computer package ‘‘ANSYSÒ [21]’’. An 8-node solid element with
three translational and additional rotational degrees of freedom
at each node was chosen to idealize the concrete and masonry
(SOLID65) whereas a 2-node bar element was used to model the
steel rebars (LINK8). Typical modeling of the column and beam
 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69 67

Fig. 30. Finite element model characterization and meshing.

400 300

300
200
200
Lateral Load (KN)

Lateral Load (KN)

100 100

0
0

-100 CLY-P-D (Exp)

CLY-S-CTRL (Exp)
-100
CLY-P-D (FEM)
-200
CLY-S-CTRL (FEM)

-300 -200

-400
-18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 -300
-35 -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 35
Top Displacement (mm)
Top Displacement (mm)
Fig. 31a. Envelope load–displacement curves for wall assembly CLY-S-CTRL.
Fig. 31c. Envelope load–displacement curves for wall assembly CLY-P-D.

300
at the top of the confined column according to same displacement
protocol of experimental tests until failure occurs. As stated before
200 from experimental results no separation was observed at the
toothed interface between the confining columns and the masonry
Lateral Load (KN)

100
0
-100
-200
-300
-20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0
Top Displacement (mm)
Fig. 31b. Envelope load–displacement curves for wall assembly CLY-P-W.
elements representing the concrete and steel rebars is indicated in
Fig. 30 with the boundary conditions. The loading of the model was
similar to that conducted in the experimental program, where a
total vertical load of 250 kN was applied uniformly on the top
beam then an incremental displacement cyclic load was applied
panel for all wall assemblies under large deformation so the inter-
face between the masonry panel and the concrete frame was mod-
eled as full bond with the corresponding mechanical properties for
each material with respect to the stress strain curves for masonry
CLY-P-W (Exp) suggested by Kaushik et al. [19].
The concrete material model in ANSYS uses a failure model
CLY-P-W (FEM)
developed by Willam and Warnke [22] for multiaxial stress state.
Solid65 element decides the cracking and crushing of concrete
through this material model. A material model may be composed
of two or more material definitions. Concrete and masonry materi-
2 4 6 8 10 12 14 16 18 20 als should have at least Elastic and Concrete material definitions. In
Elastic definition, the modulus of elasticity and Poisson’s ratio are
necessary. For Concrete definition, axial tension strength for con-
crete and masonry and shear transfer coefficients between crack
surfaces for open and closed cracks are required. If the shear trans-
fer from one crack surface to the other does not exist then the
shear transfer coefficient is 0.0, if it fully exists then the coefficient
is 1.0. In the literature, there are different suggestions for this coef-
ficient by researchers, the suggest values here were 0.3 and 0.6 for
open and closed cracks respectively for masonry as suggested by
Sandeep et al. [23].
68 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69

Table 4
Summary of finite element model vs. test results.

Wall ID Direction Maximum load (kN) Displacement relative to Cracking load (kN) Maximum displacement
maximum load (mm) (mm)
Exp. F.E.M Exp: Exp. F.E.M Exp: Exp. F.E.M Exp: Exp. F.E.M Exp:
F:E:M: F:E:M: F:E:M: F:E:M:

CLY-S-CTRL Push 300 315 0.95 7 8 0.88 250 150 1.67 16 10 1.60
Pull 280 295 0.95 10 10 1.00 170 155 1.10 14.8 10 1.48
CLY-P-W Push 250 233 1.07 12 12 1.00 155 140 1.11 16.2 12 1.35
Pull 258 242 1.07 10 12 0.83 150 136 1.10 15 12 1.25
CLY-P-D Push 205 210 0.98 15 10 1.50 140 120 1.17 27 14 1.93
Pull 209 204 1.02 20 10 2.00 120 110 1.09 22.5 14 1.61

Fig. 32a. Crack pattern for CLY-S-CTRL.

Fig. 32b. Crack pattern for CLY-P-W.

Fig. 32c. Crack pattern for CLY-P-D.

 M. El-Diasity et al. / Engineering Structures 94 (2015) 54–69
4.2. Implementation and numerical evaluation
The correlation between experimental and numerical results is
based on comparisons of failure modes, cracking patterns and plas-
tic hinge locations as well as load–displacement curves.
The predicted lateral load capacity and failure mode obtained
from the model was examined against the test results for each wall
specimens. Figs. 31a–31c show the load–displacement envelope
curves from the test and the finite element model for the solid wall
panels (CLY-S-CTRL), wall panel with window opening (CLY-P-W),
and wall panel with door opening (CLY-P-D) respectively.
Comparisons of maximum and ultimate loads with their
corresponding displacements for models and experimental results
are summarized as shown in Table 4.
The cracking patterns for the finite element models for each
wall assembly as compared to the experimental test results are
shown in Figs. 32a–32c.
Results from the finite element analysis of showed that the
developed models are capable with sufficient degree of accuracy
to capture the maximum load and its corresponding deformation
of the tested walls except for corresponding deformation for wall
CLY-P-D. The ultimate displacements in models were less than its
in experimental works which indicated that the plastic analysis
needs to be enhanced.
The proposed model showed good agreement with the results
of the laboratory tests for crack patterns and failure mechanisms
for all models as shown in Figs. 32a–32c.
5. Summary and conclusions
This paper presents results of cyclic loading tests investigating
the in-plan behavior of confined masonry walls retrofitted using
low cost ferrocement and GFRP. Ten walls with scale of 0.8 were
built, using 0.8 scale brick clay units, consisting of a clay masonry
panel, two confining columns and a tie beam, were tested under a
combination of a constant vertical load and lateral cyclic loading
with displacement control protocol up to failure. Wall panels had
various configurations, namely, six solid walls, two perforated
walls with window openings and two with door openings. Two
composite materials (ferrocement and GFRP) and three retrofitting
configurations (diagonal ‘‘X’’, connection, and full coverage) were
investigated. Key research findings may be summarized as follows:
1. The upgrading techniques improved the lateral resistance of the
confined walls by a factor ranging from 25% to 32% and also
improve the total energy dissipation by a factor ranging from
33% to 85%.
2. The upgrading configurations of diagonal ‘‘X’’ and full coverage
can help prevent diagonal shear failure especially in tie columns
and convert failure mode to rocking mode for the undamaged
masonry panel. Additionally, in all situations, it will postpone
collapse by ‘‘keeping the bricks together’’ under large seismic
deformations.
3. The improvement in maximum lateral drifts was less
significant.
4. Ferrocement and the low cost GFRP sheets showed a similar
enhancement for solid panels without any de-bonding under
high levels of lateral loading.
5. The proposed finite element models showed good agreement
with the results of the laboratory tests for crack patterns and
failure mechanisms for all models and for maximum load ant
its corresponding deformation for most cases.
69
6. The proposed finite element models showed good agreement
with the results of the laboratory tests for maximum load and
its corresponding deformation for most cases.
References
[1] Brzev S. Earthquake-resistant confined masonry construction. India: National
Information Center of Earthquake Engineering, Indian Institute of Technology
Kanpur; 2008.
[2] Yoshimura K, Kikuchi K, Kuroki M, Nonaka H, Tae Kim K, Wangdi R, Osikata A.
Experimental study for developing higher seismic performance of brick
masonry walls. In: 13th World conference on earthquake engineering,
Vancouver, B.C., Canada, No. 1597; 2004.
[3] Riahi Zahra, Elwood Kenneth J, Alcocer Sergio M. Backbone model for confined
masonry walls for performance-based seismic design. J Struct Eng
2009;135(6):644–54.
[4] Tomazevic M, Klemence I. Verification of seismic resistance of confined
masonry buildings. Earthquake Eng Struct Dyn 1997;26:1073–88.
[5] Tomazevic M. Earthquake-resistant design of masonry buildings. London,
U.K: Imperial College Press; 1999.
[6] Tu YH, Chuang TH, Liu PM, Yang YS. Out-of-plane shaking table
tests on unreinforced masonry panels in RC frames. Eng Struct
2010;32:3295–935.
[7] Ishibashi K, Meli R, Alcocer SM, Leon F, Sanchez TA. Experimental study on
earthquake-resistant design of confined masonry structure. In: Proceedings of
the tenth world conference on earthquake engineering, Madrid, Spain; 1992. p.
3469–74.
[8] Gostic S, Zarnic R. Cyclic lateral response of masonry infilled RC frames and
confined masonry walls. In: Proceedings of the 8th North American masonry
conference, Austin; 3–6 June 1999. p. 477–88.
[9] Yanez F, Astroza M, Holmberg A, Ogaz O. Behavior of confined masonry shear
walls with large openings. In: 13th World conference on earthquake
engineering, Vancouver, B.C., Canada, No. 343; 2004.
[10] Lourdes Ana, Tomohisa. Evaluation of structural performance for confined
masonry walls retrofitted with wire mesh and with disposable fiber mats. In:
Proceedings of the 44th International Institute of Seismology and Earthquake
Engineering (IISEE), Peru; 2009. p. 97–102.
[11] Flores LE, Alcocer SM. Calculated Response of confined masonry structures. In:
11th World conference on earthquake engineering, Mexico, paper No. 1830;
1996.
[12] Yoshimura K, Kikuchi K, Kuroki M, Nonaka H, Tae Kim K, Wangdi R, Oshikata A.
Experimental study on effect of height of lateral forces, column reinforcement
and wall reinforcement on seismic behavior of confined masonry walls. In:
13th World conference on earthquake engineering, Vancouver, Canada, No.
1597; 2004.
[13] ElGawady MA, Lestuzzi P, Bandoux M. In-plane seismic response of URM walls
upgraded with FRP. J Compos Constr 2005;2005:524–35.
[14] Yu P, Silva P, Nanni A. In-plane response of URM walls strengthened with GFRP
grid reinforced polyurea. In: 10th North American masonry conference, USA;
2007. p. 466–77.
[15] ASTM E519-02, 2002. Standard test method for diagonal tension (shear) in
masonry assemblages.
[16] European Committee for Standardisation, EC6, Design of masonry structures.
Part 1-1: general rules for buildings – rules for reinforced and un-reinforced
masonry, ENV1996 1-1: Bruxels: CEN; 1995.
[17] RILEM TC: 76-LUM. Diagonal tensile strength tests of small wall specimens. In:
RILEM (International Union of Laboratories and Experts in Construction
Materials), recommendations for the testing and use of constructions
materials, London: E&FN SPON; 1994. p. 488–89.
[18] Hose Y., Seible F. Performance evaluation database for concrete bridge
components, and systems under simulated seismic loads. PEER report 1999/
11, Pacific Earthquake Engineering Research Center College of Engineering,
University of California, Berkley, U.S.A.; 1999.
[19] Kaushik Hemant B, Rai Durgesh C, Jain Sudhir K. Stress–strain characteristics
of clay brick masonry under uniaxial compression. J Mater Civ Eng
2007;19(9):728–39. ASCE.
[20] Huang Yan, Kan Minghui, Wang Zifa. Nonlinear analysis for monotonic and
low cyclic loading. Appl Mech Mater 2011;94–96:406–15.
[21] ANSYSÒ Academic Research, Release 12.0, ANSYS Inc.
[22] Willam KJ, Warnke ED. Constitutive model for the triaxial behavior of concrete.
In: Proceedings of International Association for Bridge and Structural
Engineering: seminar on concrete structures subjected to triaxial stress,
ISMES, Bergamo, Italy; 1975. p. 174186.
[23] Sandeep, Renukadevi MV, Manjunath S, Somanath. Influence of reinforcement
on the behavior of hollow concrete blocks masonry prism under compression
– an experimental and analytical approach. Int J Res Eng Technol
2013:106–10.